Unveiling multiscale dynamics: exploring financial risk spillover and influencing factors among Chinese financial institutions

This paper aims to uncover patterns of financial risk spillover between Chinese financial institutions and investigate influencing factors across various timescales and periods. First, we decompose the transaction prices of financial institutions using the maximal overlap discrete wavelet transform (MODWT). Using the copula model, we construct multiscale, multiperiod interconnected networks between Chinese financial institutions at different timescales. Then, we measure risk spillover between financial institutions using the dynamic Delta-conditional value-at-risk (∆CoVaR) model. Subsequently, we integrate the correlation matrixes into a spatial econometric model to explore the influencing factors of risk spillover between financial institutions at different timescales and in different periods using samples of 27, 42 and 78 financial institutions listed in China from 2008 to 2023. Our results highlight the critical role of banking and securities institutions in transmitting network risk, although the importance of banks is found to decrease over time. During financial events, the clustering trend within the banking industry decreases. A rise in stock market returns, the size of financial institutions, market-to-book ratios or financial leverage increases the risk spillover to other related financial institutions. Conversely, rising asset returns lead to a reduction in risk spillover to other related financial institutions. Based on these findings, we propose recommendations for preventing financial risks.